INDEX

Explanations

terms related to physical arms or weapons

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

light

-0.68

ples

-0.61

ters

-0.61

DER

-0.58

PER

-0.57

charged

-0.57

rers

-0.56

faced

-0.56

TER

-0.55

flush

-0.55

POSITIVE LOGITS

ament

1.17

aments

1.13

aceutical

1.06

ando

1.05

ageddon

1.02

chair

0.96

agnetic

0.93

ally

0.93

heid

0.91

ophon

0.91

Activations Density 1.431%

terms related to physical arms or weapons

No Comments

No Known Activations